Method and apparatus for increasing scavenging garbage collection effectiveness

Abstract

Methods and apparatus for enabling an efficient generational scavenging garbage collection to be performed on a managed memory area are disclosed. According to one aspect of the present invention, a method for reclaiming memory space uses a managed memory area that includes a first area and a second area. The first area is arranged to store recently allocated objects, while the second area being arranged to store older objects. The method includes determining when a first section of the first area in which new objects are to be allocated is substantially filled. When it is determined that the first section is substantially filled, a garbage collection is performed on a second section of the first. After the garbage collection, the second section is set to support new object allocation such that new objects are allocated in the second section, while the first section is reset such that it is no longer arranged to support new object allocation. In one embodiment, performing the garbage collection on the second section includes copying a live object from the second section into the second area.

Description

1. Field of InventionThe present invention relates generally to methods and apparatus for improving the performance of software applications. More particularly, the present invention relates to methods and apparatus for reducing the overhead associated with performing memory allocation.2. Description of the Related ArtThe amount of available memory is typically limited in computing systems such as object-based computing systems. Hence, memory must generally be conserved and recycled. Many computer programming languages enable software developers to dynamically allocate memory within a computer system. Some programming languages require explicit manual deallocation of previously allocated memory, which may be complicated and prone to error. Languages which require explicit manual memory management include the C and C++ programming languages. Other programming languages utilize automatic storage-reclamation to reclaim memory that is no longer necessary to ensure the proper operation o...

AI Technical Summary

Problems solved by technology

The amount of available memory is typically limited in computing systems such as object-based computing systems.

Some programming languages require explicit manual deallocation of previously allocated memory, which may be complicated and prone to error.

However, if new generation object 110d is dead, a garbage collection performed in new generation 104 may result in old generation object 110f becoming unreachable if old generation object 110f is not be pointed to by any other object.

In general, copying operations are slow and expensive.

As a result, the overall garbage collection process is expensive.

A scavenging garbage collection is often an expensive process.

Specifically, as mentioned above, the copying of a live object during a scavenging garbage collection is a slow and expensive operation.

Hence, when many live objects are copied during a scavenging garbage collection, or when a few large objects are copied during a scavenging garbage collection, the garbage collection process itself becomes both slow and costly.

Images